Robustness
Robustness is the property that any post-crash states of the program under weak persistency model is also an observable post-crash state if the program is executed in Strict Persistency.
Robustness is a sufficient but not necessary condition for Crash Consistency (more specifically, for lock-free programs only1). This is because programs may (1) be tolerant of reading stale values, e.g., counters that only need to be approximately correct, or (2) use other mechanisms like checksums to detect and discard inconsistent data after reading it.
It is proposed by Hamed Gorjiara, Weiyu Luo, Alex Lee, Guoqing Harry Xu, and Brian Demsky in the PLDI’22 paper Checking Robustness to Weak Persistency Models. See Checking Robustness to Weak Persistency Models